FrigoChef
Completed academic team project
What it was
FrigoChef is a native Android application that recommends recipes based on the ingredients and quantities available at home. Built in Kotlin by a three-person team, it combines a guided questionnaire, recipe filtering, compatibility scoring, SQLite persistence and automated tests.
What I owned
My contribution focused on the database layer, the results and recipe-detail screens, code reviews, and unit and end-to-end testing.
The hard part
The main challenge was translating ingredient quantities into a score that remained simple and explainable. For each required ingredient, the application calculates the available-to-required quantity ratio, caps it at 100%, assigns zero when the ingredient is missing, and averages the results across the recipe. This produces a proportional score rather than a binary match, and distinguishes a missing ingredient from a partially available one.
The calculation runs in Kotlin, in a dedicated ScoreCalculateur. Every ingredient currently carries the same weight, which is a deliberate simplification: a recipe where I have 3 of 3 ingredients scores 100% and ranks above one where I have 10 of 12 at 83%. Tests cover the zero, partial and complete cases.
What I would do differently
In a future version, I would distinguish essential ingredients from optional ones, move database work to Room with coroutines, and test the scoring model against a larger and more varied recipe catalogue.
